Innovations in Biotechnology Part 16 potx

Recovery and purification processes vary widely depending on the manufacturing process and specific mAb characteristics, but generally, the isolation and purification of mAbs involve a c

environment for days to weeks, and then the cell culture fluid is harvested by centrifugation (Shukla & Kandula, 2008) In mammalian cells, the product monoclonal antibodies are secreted from the cells into the supporting fluid medium Centrifugation separates the cells from the fluids and facilitates simpler recovery procedures downstream

Commercial mAb production requires considerable preproduction effort to ensure that the cell line is stable and can produce commercially appropriate quantities of antibody In addition, the commercial production process must produce a product that meets the quality expectations of regulatory authorities In the past few years, improvements have been made in critical areas, such as cell line generation and large-scale cell culture production, to maximize specific antibody productivity from a given cell line and improve overall productivity in bioreactors These advances include the use of new expression vectors and transfection technology, high-throughput, robust screening technologies to select the highest producing clones rapidly and more effectively, improvements in cell culture and optimized bioreactor processes (Li et al., 2010; Schlatter et al., 2005) As a result, the production of cell lines expressing multigram quantities of antibody per liter of culture medium is now routine The product quality and product heterogeneity of every mAb is highly dependent on its manufacturing process (Abu-Absi et al., 2010; Horvath et al., 2010) The ideal manufacturing conditions would have optimal production levels of product in conjunction with the desired product quality profile Attributes that are typically deemed critical in selecting stable clones and cell culture conditions are the product titer and product heterogeneity, including charged species and aggregates Production titers directly correlate to the costs of the process and are desired to be as high as possible with minimal impact to other quality attributes of the product (Kelley, 2009) Critical quality attributes of the product, such as the level of aggregation, are carefully monitored, as failure to control critical quality attributes may pose a safety risk to the patient (Rosenberg, 2006)

4 Monoclonal antibody purification and formulation

Once monoclonal antibodies are produced in cells, the mAbs must be recovered and purified Recovery and purification processes vary widely depending on the manufacturing process and specific mAb characteristics, but generally, the isolation and purification of mAbs involve a centrifugation step to separate the cells from the cell culture fluid containing the mAb product, one or more chromatography steps, which can include affinity chromatography, cation or anion exchange chromatography, hydrophobic interaction chromatography (HIC) and displacement chromatography (Shukla et al., 2007), and

filtration or precipitation steps (Gottschalk, 2009) Many of the purification steps are designed to remove contaminants and adventitious agents (e.g., bacteria, fungi, viruses, and mycoplasma)

After elution from the final chromatographic purification step, a unit operation is required

to exchange the components of the chromatography elution buffer with the chosen formulation components The predominant technology that has been used in the industry for buffer exchange and concentration is ultrafiltration/diafiltration using tangential-flow filtration (Genovesi, 1983; Shiloach, 1988; van Reis, 2001) After this step, the drug substance

is filtered and typically frozen as bulk for storage until filling occurs to produce the final drug product

The formulation of the mAb therapeutic is chosen in part to ensure product quality during shelf life Formulations are designed to minimize protein aggregation, decrease viscosity, and increase shelf life through preventing degradation (Shire, 2009) High protein concentration formulations are being developed to allow for subcutaneous or intramuscular delivery of mAb products (Shire et al., 2004) Historically, the most conventional route of delivery for protein drugs has been intravenous administration because of poor bioavailability by most other routes, greater control during clinical administration, and faster pharmaceutical development Subcutaneous delivery allows for home administration and improved patient compliance However, development of high protein concentration formulations involves unique manufacturing challenges compared to low concentration formulations, such as higher viscosities and necessary changes to unit operation steps

5 Monoclonal antibody characterization and release testing

Biopharmaceutical manufacturing of monoclonal antibodies produces a heterogeneous product of structurally related species Antibody speciation can occur throughout the manufacturing process at various steps, including cell culture, harvest, purification, formulation, filling and during shelf life Full-length monoclonal antibodies are high molecular weight proteins (around 150,000 Da), and have highly complex secondary and tertiary structures, subject to post-translational modifications Therefore, product characterization and quality control testing are required at critical points throughout clinical development and manufacturing to control for these species (Harris et al., 2004) Figure 2 depicts the structure of a monoclonal antibody compared to a small molecule drug, illustrating the increased complexity of a biologic compared to a small molecule therapeutic Antibodies can be characterized by many physicochemical properties including hydrated size (Stokes radius), molecular weight, charge, hydrophobicity, electrophoretic mobility, isoelectric point (pI), sedimentation velocity, glycosylation, and spectral properties The nature of each species can be related to differences in their primary, secondary, tertiary, or quaternary protein structures In addition, monoclonal antibodies are susceptible to chemical or enzymatic modification, particularly at sites that are exposed to the protein-liquid interface Product heterogeneity can be caused by a number of modifications, such as C-terminal processing of lysine residues (Harris, 1995; Santora et al., 1999; Weitzhandler et al., 1998), deamidation (Di Donato et al., 1993; Hsu et al., 1998), glycation (nonenzymatic glucose addition) (Quan et al., 2008), amino acid sequence variations (Yang et al., 2010), and noncovalent complexes (Santora et al., 2001)

molecule will be well-characterized before the Biologic License Application (BLA) is submitted to the regulatory agencies

Fig 2 Comparison of the structures of a mAb (Herceptin) and a small molecule therapeutic (Tarceva)

Many of the recommended protein characterization assays are based on liquid chromatography methods, such as ion exchange chromatography (IEC) for charge heterogeneity analysis, size exclusion chromatography (SEC) for size heterogeneity, and reversed-phase high performance liquid chromatography (RP-HPLC) for peptide mapping (Chirino & Mire-Sluis, 2004) The remainder of this chapter will primarily focus on ion exchange chromatography methods for analyzing charge heterogeneity for characterization and support of formulation and process development, as well as for lot release testing of drug substance and drug product (Schnerman et al., 2004)

5.1 Analyzing mAb charge heterogeneity using IEC

As mentioned previously, monoclonal antibodies are large proteins that are quite complex While the light chain and heavy chain sequences of a particular mAb may be known, a number of modifications can introduce heterogeneity in the product Thus, it is important to develop appropriate analytical methods to resolve the minor forms of the product Analytical biochemists routinely use IEC for resolving charge variants of the protein The scientist must then utilize orthogonal analytical methods to characterize the separated peaks

of the ion exchange chromatogram The characterization of a mAb is particularly important

if the modifications occur in the complementarity-determining regions (CDR), as modifications in the CDR can affect the binding activity and potency of the mAb

A strategy for the assignment of peaks from a weak cation exchange (WCX) mAb separation using a salt gradient has been published (Harris et al., 2001) Seven forms of a therapeutic recombinant antibody were resolved by cation-exchange chromatography The peak fractions were collected, and structural differences were assigned by peptide mapping, which involves digesting the mAb with an enzyme and injecting the digest onto

a reverse-phase column coupled to a mass spectrometer, and by hydrophobic interaction chromatography (HIC) after papain digestion The peaks in this particular case were attributed to deamidation, isomerization, and succinimide intermediates Other orthogonal analytical methods were used to characterize the IEC peaks; one of these methods—potency testing—determined that one minor peak demonstrated much lower potency than the main peak

In another study, a recombinant humanized monoclonal IgG1 antibody with different states of glycosylation on the conserved asparagine residue in the CH2 domain was analyzed by cation exchange chromatography (Gaza-Bulseco et al., 2008) Two major peaks were observed and were further characterized by enzymatic digestion and mass spectrometry It was found that this recombinant monoclonal antibody contained three glycosylation states—zero, one or two glycosylated heavy chains The peak that eluted earlier on the cation exchange column contained antibodies with two glycosylated heavy chains containing fucosylated biantennary complex oligosaccharides with zero, one or two terminal galactose residues The peak that eluted later from the column contained antibodies with zero, one or two glycosylated heavy chains The oligosaccharide on the antibodies that eluted in the later peak was composed of only two GlcNAc residues These results indicate that conformational changes, caused by different types of neutral oligosaccharides as well as the absence of certain oligosaccharides, can be differentiated

by cation exchange column chromatography

5.2 Lot release testing of mAbs

Once the mAb is purified and formulated, the resulting drug substance must be tested prior

to lot release A set of tests and acceptance criteria are established based on mAb characterization and regulatory requirements in order to ensure product quality (Food & Drug Administration (FDA), 1999) These tests typically include appearance, identity, purity, protein concentration, potency of the molecule, microbial limits or bioburden, and bacterial endotoxins (Table 1) IEC is one of the most frequently used lot release methods for purity for mAbs (Schnerman et al., 2004) Once these tests are performed and the results

Purity

Limulus Amebocyte Lysate (Endotoxin) Size Exclusion Chromatography (SEC) CE-SDS (Capillary Electrophoresis-Sodium Dodecyl Sulfate) IEC (Ion Exchange Chromatography) or icIEF (Imaged Capillary

Isoeletric Focusing) Glycosylation Profile Peptide Mapping by RP-HPLC

General Tests

Osmolality

pH Surfactant Concentration (e.g Polysorbate 20) Table 1 Commonly used tests found on a Certificate of Analysis for lot release; a selected

subset is used for stability testing of mAbs

6 Mechanism of ion exchange chromatography of mAbs

Ion exchange chromatography (IEC) has been a platform for monoclonal antibody

purification and characterization for many years For the analysis of charged species of

proteins, IEC is a popular method due to the fact that it preserves the native conformation

and maintains bioactivity of the protein, is relatively easy of use, is supported by the

maturity of the equipment and consumables market, and has widespread use in the

biopharmaceutical industry (Rea et al., 2010)

Charge-based methods are an integral component of characterization studies and quality

control strategies because they are sensitive to many types of modifications Charge

profiling of intact antibodies can resolve species related to protein conformation, size,

sequence species, glycosylation and post-translational modifications (Gaza-Bulseco et al.,

2008; Harris et al., 2001; He et al., 2010) Although IEC can be used to track specific species, it

is common to group all species not associated with the main peak and report them as either

acidic or basic species (Figure 3) In addition, fractions collected from an IEC run can often

be directly injected onto orthogonal columns for further analysis, such as reverse-phase and

size exclusion chromatography columns, or submitted for potency testing

IEC separates proteins based on differences in the surface charge of the molecules, with separation being dictated by the protein interaction with the stationary phase The two main categories of ion exchange chromatography are cation exchange (CEX) and anion exchange (AEX) Cation exchange chromatography retains biomolecules by the interaction of the negatively-charged resin with histidine (pK ~ 6.5), lysine (pK ~ 10) and arginine (pK ~ 12) in the protein Anion exchange chromatography primarily retains biomolecules by the interaction of the positively-charged resin with aspartic or glutamic acid side chains, which have pKa of ~4.4 In addition to the amino acid residues, cation exchange columns can also separate deamidated, glycated and other charged variants Anion exchange columns have also been useful for separating phosphorylated and hydroxyl modified amino acids When the pH equals the pI value of the protein, the net charge on the molecule is zero However, significant retention can occur for proteins even when the pH of the mobile phase is equal to the pI of the molecule; despite an overall net charge of zero, only a portion of the mAb molecule will interact with the stationary phase, and there will be a net charge on that portion of the molecule because

of an uneven distribution of charged groups throughout the molecule (Vlasak & Ionescu, 2008) Thus, it is possible to separate proteins having very similar charge (Figure 4), or even structural isomers with identical pI values, by ion exchange chromatography

Fig 3 Typical cation exchange chromatogram for analytical characterization of a mAb Integration is shown, and main peak, acidic and basic regions are denoted

There are two ways to elute the protein from the IEC column: 1) increasing salt concentration with time or 2) by varying the mobile phase pH value as a function of time Increasing the salt concentration elutes the protein by increasing the ionic strength of the mobile phase, thus affecting the charge interaction of the mAb and the stationary phase A

pH gradient elutes the protein by changing the charge on the molecule, thus affecting the binding of the molecule to the stationary phase While conventional salt gradient cation exchange chromatography is regarded as the gold standard for charge sensitive antibody analysis (Vlasak & Ionescu, 2008), method parameters such as column type, mobile phase

pH, and salt concentration gradient often need to be optimized for each individual antibody

A recent publication described a multi-product pH gradient IEC method for the separation

of mAb charge species for a variety of mAbs using a single method (Farnan & Moreno, 2009) The following sections will discuss both salt-gradient and pH-gradient based elution methods, and the combination of the two modalities (hybrid methods)

Fig 4 Separation of mAbs differing by only one charge, a single amino acid change to primary structure The elution buffer (0.5 M NaCl in 20 mM Tris, pH 7.3) was increased linearly on a ProPac WCX-10 column (4 x 250 mm), which was held at 50 ºC and had a flow rate of 1 mL min-1

7 Developing a salt-based IEC method

Salt-based IEC separations are developed by choosing a cation or anion exchange column and varying the buffer system, mobile phase pH value, and ionic strength gradient of the elution buffer Figure 5 shows a typical development workflow for salt-based IEC and pH-based IEC development, and can serve as a guide for initial IEC method development The following sections will cover in more detail the outputs to consider when screening various parameters during development More general considerations regarding HPLC method development can be found in various texts (Kastner, 2000; Snyder et al., 1997)

Fig 5 Sequential salt-gradient IEC and pH-gradient IEC method development and

optimization work flow

7.1 Column selection, buffers and operating parameters for salt gradient IEC

Column selection is perhaps the most subjective part of the optimization process; picking between the different vendor offerings and functionalities can be difficult Prior experience, data in the literature or unpublished results within the organization are often the best starting points

Analytical ion exchange chromatography of proteins is typically carried out using mobile phases that are relatively neutral in pH values, 5.5 to 8.5 This general practice is recommended because at pH extremes, the protein is more likely to degrade The selection

of whether to use anion or cation exchange chromatography is also driven by the isoelectric point of the protein (pI) and the species to be resolved, e.g., phosphorylated species, C-terminal lysine variants, etc

If the pI value of the mAb is greater than 8, a CEX column is evaluated at pH 6-7 initially CEX primarily retains mAbs by the interaction of acid groups on the CEX resin with lysine, arginine and histidine side chains on the mAb Since mAbs are positively charged at a mobile phase pH below their pI, the mAb species would likely be retained and resolved on a CEX column under the recommended mobile phase pH range

If the pI value of the mAb is less than 6, an AEX column is evaluated at a pH above 6 initially AEX primarily retains biomolecules by the interaction of amine groups on the ion exchange resin with aspartic or glutamic acid side chains Since mAbs are negatively charged at a mobile phase pH above their pI, the mAb species would likely be retained and resolved on an AEX column

For intermediate pI values of 6-8, both CEX and AEX are evaluated because of the possibility that the portion of the mAb that interacts with the stationary phase, typically the side chains that are exposed to the mobile phase, has a different charge than the pI would suggest, e.g., the surface charge of the mAb is positive despite the entire mAb having an overall negative charge Ultimately, the species of interest that are to be resolved determine whether CEX or AEX is chosen for molecules with intermediate pI’s; the separation mode that better separates the species of interest is usually the one that is chosen for mAb analysis Figure 6 shows CEX and AEX chromatograms of a Fab (mAb fragment) reference sample and thermally stressed sample In this case, the Fab molecule has a nominal pI value for the main species of 7.6 It should be noted that the separations on the AEX and CEX columns were each optimized independently for column type, pH value and salt gradient It should also be noted that the terms “strong” and “weak” (in SAX, strong anion exchange, and WCX, weak cation exchange) refer to the extent of variation of ionization with pH due to the functional groups on the resin and not the strength of binding Strong ion exchangers are completely ionized over a wide pH range whereas with weak ion exchangers, the degree of dissociation and thus exchange capacity varies much more markedly with pH For this example, SAX results in significantly more peaks and much better resolution of the charge species in comparison to the WCX chromatogram Particularly interesting is that the difference between the WCX and SAX elution profiles are much more vivid for the stressed samples than for the reference materials We have seen examples where the converse is true and the CEX separation is better than that observed on the AEX This contrast between the AEX and CEX profiles highlights an important feature of IEC that electrophoretic methods don’t exhibit, which is the ability to magnify particular aspects of the protein structure and accentuate the separation of species relating to particular motifs (Vlasak & Ionescu, 2008)

Fig 6 Separation of Fab charge species using a weak cation exchange column (WCX) and a strong anion exchange column (SAX) Thermally stressed samples are labeled by incubation time and temperature of incubation

In general, we have observed that for the separation of mAb variants using ion-exchange chromatography, the optimized chromatogram has a relatively shallow gradient over a narrow range of salt concentration A typical method results in 100 mM NaCl as the center point of the gradient, with salt concentration increasing over 70 mM NaCl in a linear gradient It is recommended to perform iterative gradient optimizations to narrow the NaCl gradient down to around 2 mM/mL min-1 Iterative cycles are quicker and more predictive than performing a very long shallow gradient

Chromatograms obtained during the mobile phase pH value optimization for a mAb with a

pI value around 9.5 are shown in Figure 7 Buffer species and buffer concentration for gradient IEC are generally not significant factors, but should be chosen considering target

salt-pH and buffer pKa

Although temperature does not significantly affect electrostatic interactions, it often affects the pH value of the mobile phase This is particularly of concern for a Good’s buffer system (group of buffers described in the research of Dr Norman Good et al in 1966, often used for IEC and other biochemistry applications) (Good et al., 1966), which can exhibit a change in

pH value of around 0.02 per °C temperature change This sensitivity creates a need to control the column temperature carefully A column compartment is always used, typically set at a value greater than 30°C to ensure good temperature stability in compartments that can only apply heat Above 30°C, temperature control within +/- 1°C is readily achievable with commercially available equipment

Fig 7 Effect of mobile phase pH on mAb separation by WCX The elution buffer (0.5 M NaCl) was increased linearly at 1 mM min-1 at a flow rate of 1 mL min-1 on a ProPac WCX-10 column (4 x 250 mm), which was held at 30 ºC Different initial salt concentrations were optimized for each pH value Integration is shown, and main peak, acidic and basic regions are denoted

Subtle variations in selectivity with temperature may result from temperature-induced changes in mobile phase pH value (Figure 8) In Figure 8, the elution profile changes in two distinct regions as a function of temperature Below 40°C, subtle changes in elution profile and retention times are observed consistent with minor changes to the mobile phase pH value as a function of temperature However, above 40°C, the profiles exhibit much more radical changes with increasing temperature This is interpreted to be related to the mAb having lost higher order structure at those elevated temperatures due to protein denaturing For the mAb in Figure 8, it is clear that moderately elevated temperatures are not possible while maintaining the higher order structure; in general for IgG1 mAbs, chromatography at temperatures up to 55°C is readily possible In summary, while mobile phase temperature does not affect protein charge directly, temperature can affect mobile phase pH and the structure of the protein, which can affect chromatographic separations Thus, column temperature should be optimized considering these temperature effects

Fig 8 Effect of temperature on mAb separation by CEX The elution buffer (0.2 M sodium sulfate) was increased linearly on a ProPacWCX-10 column (4 x 250 mm)

8 Developing a pH gradient-based IEC method

Despite good resolving power and robustness, salt-based ion exchange separations are usually protein-specific and time-consuming to develop A novel pH-based separation of proteins by cation exchange chromatography that was multi-product, high-resolution, and robust against variations in sample matrix salt concentration and pH was recently reported (Farnan & Moreno, 2009) A pH gradient-based separation method using cation exchange chromatography was also evaluated in a mock validation and deemed highly robust (Rea et al., 2011) Figure 9 depicts the separation of 16 mAbs by pH gradient IEC (pH-IEC) Each mAb was injected sequentially, demonstrating that in contrast to salt-based IEC, pH-IEC can

be used to analyze multiple mAbs with a single method

Fig 9 Separation of 16 mAbs using a ProPac WCX-10 column by pH gradient IEC Each mAb was analyzed using the same pH-IEC method, and each mAb was injected

sequentially mAb pI values ranged from pI 7.3 to pI 9.4

Similar to salt-gradient IEC methods, pH-IEC separations are developed by choosing a cation or anion exchanging column and varying the buffer system, pH of the mobile phases, and other operating parameters, such as temperature and flow rate Figure 5 shows a typical development workflow for pH-IEC, and can serve as a guide for initial pH-IEC method development

8.1 Column selection, buffers and operating parameters for pH gradient IEC

Like conventional IEC, the conditions chosen for pH-IEC separations, such as buffer, pH, column temperature, and sample load, are dependent on the type of column selected To choose a column, the pI of the mAb and the expected charge species should be considered Considerations for column selection may differ slightly for pH-IEC compared to conventional IEC For example, because the column will be exposed to a pH gradient, the column must be able to perform adequately over a large pH range, i.e., the charged groups on the chromatography resin must maintain their charge over the operating pH range Also, buffer strength can affect resolution, and pH-IEC mobile phases typically have lower buffer strengths than conventional salt-gradient IEC Several pH-IEC buffer systems have been published for mAb separations; these buffer systems can be used as starting points for formulating buffers for pH-IEC methods (Farnan and Moreno, 2009; Rea et al., 2011; Rozhkova, 2009)

8.2 High-throughput multi-product separations using pH-IEC

To increase the throughput of the analytical methods, smaller particle sizes and shorter column lengths are being utilized to reduce run time In Figure 10, the utilization of a 3 µm particle size column reduced analysis time 16-fold compared to a 10 µm particle size column Analysis times are greatly reduced using smaller particle sizes because as the particle size decreases, there a significant gain in column efficiency, and the efficiency does not decrease at increased flow rates or linear velocities (Swartz, 2005) In addition, because different mAbs can be analyzed using the same pH-IEC method in the same sequence, these high-throughput methods are capable of analyzing hundreds of mAbs per day, which is not possible with conventional, product-specific salt-based IEC

Fig 10 Separation of a mAb using (A) a WCX column, 10 µm and (B) a SCX column, 3 µm,

by pH gradient IEC Each mAb was analyzed using the same buffers and gradient volume